netdev - Re: [PATCH bpf-next 2/2] bpf: Add test for SO_REUSEPORT_DETACH

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20190612214757.GC9056@mini-arch>
Date:   Wed, 12 Jun 2019 14:47:57 -0700
From:   Stanislav Fomichev <sdf@...ichev.me>
To:     Martin Lau <kafai@...com>
Cc:     "bpf@...r.kernel.org" <bpf@...r.kernel.org>,
        "netdev@...r.kernel.org" <netdev@...r.kernel.org>,
        Alexei Starovoitov <ast@...com>,
        Daniel Borkmann <daniel@...earbox.net>,
        David Miller <davem@...emloft.net>,
        Kernel Team <Kernel-team@...com>
Subject: Re: [PATCH bpf-next 2/2] bpf: Add test for SO_REUSEPORT_DETACH_BPF

On 06/12, Martin Lau wrote:
> On Wed, Jun 12, 2019 at 12:59:17PM -0700, Stanislav Fomichev wrote:
> > On 06/12, Martin KaFai Lau wrote:
> > > This patch adds a test for the new sockopt SO_REUSEPORT_DETACH_BPF.
> > > 
> > > '-I../../../../usr/include/' is added to the Makefile to get
> > > the newly added SO_REUSEPORT_DETACH_BPF.
> > > 
> > > Signed-off-by: Martin KaFai Lau <kafai@...com>
> > > ---
> > >  tools/testing/selftests/bpf/Makefile          |  1 +
> > >  .../selftests/bpf/test_select_reuseport.c     | 50 +++++++++++++++++++
> > >  2 files changed, 51 insertions(+)
> > > 
> > > diff --git a/tools/testing/selftests/bpf/Makefile b/tools/testing/selftests/bpf/Makefile
> > > index 44fb61f4d502..c7370361fa81 100644
> > > --- a/tools/testing/selftests/bpf/Makefile
> > > +++ b/tools/testing/selftests/bpf/Makefile
> > > @@ -16,6 +16,7 @@ LLVM_OBJCOPY	?= llvm-objcopy
> > >  LLVM_READELF	?= llvm-readelf
> > >  BTF_PAHOLE	?= pahole
> > >  CFLAGS += -Wall -O2 -I$(APIDIR) -I$(LIBDIR) -I$(BPFDIR) -I$(GENDIR) $(GENFLAGS) -I../../../include \
> > > +	  -I../../../../usr/include/ \
> > Why not copy inlude/uapi/asm-generic/socket.h into tools/include
> > instead? Will that work?
> Sure. I am ok with copy.  I don't think we need to sync very often.
> Do you know how to do that considering multiple arch's socket.h
> have been changed in Patch 1?
No, I don't know how to handle arch specific stuff. I suggest to copy
asm-generic and have ifdefs in the tests if someone complains :-)

> Is copy better?
Doesn't ../../../../usr/include provide the same headers we have in
tools/include/uapi? If you add -I../../../../usr/include, then is there
a point of having copies under tools/include/uapi? I don't really
know why we keep the copies under tools/include/uapi rather than including
../../../usr/include directly.

> > >  	  -Dbpf_prog_load=bpf_prog_test_load \
> > >  	  -Dbpf_load_program=bpf_test_load_program
> > >  LDLIBS += -lcap -lelf -lrt -lpthread
> > > diff --git a/tools/testing/selftests/bpf/test_select_reuseport.c b/tools/testing/selftests/bpf/test_select_reuseport.c
> > > index 75646d9b34aa..5aa00b4a4702 100644
> > > --- a/tools/testing/selftests/bpf/test_select_reuseport.c
> > > +++ b/tools/testing/selftests/bpf/test_select_reuseport.c
> > > @@ -523,6 +523,54 @@ static void test_pass_on_err(int type, sa_family_t family)
> > >  	printf("OK\n");
> > >  }
> > >  
> > > +static void test_detach_bpf(int type, sa_family_t family)
> > > +{
> > > +	__u32 nr_run_before = 0, nr_run_after = 0, tmp, i;
> > > +	struct epoll_event ev;
> > > +	int cli_fd, err, nev;
> > > +	struct cmd cmd = {};
> > > +	int optvalue = 0;
> > > +
> > > +	printf("%s: ", __func__);
> > > +	err = setsockopt(sk_fds[0], SOL_SOCKET, SO_DETACH_REUSEPORT_BPF,
> > > +			 &optvalue, sizeof(optvalue));
> > > +	CHECK(err == -1, "setsockopt(SO_DETACH_REUSEPORT_BPF)",
> > > +	      "err:%d errno:%d\n", err, errno);
> > > +
> > > +	err = setsockopt(sk_fds[1], SOL_SOCKET, SO_DETACH_REUSEPORT_BPF,
> > > +			 &optvalue, sizeof(optvalue));
> > > +	CHECK(err == 0 || errno != ENOENT, "setsockopt(SO_DETACH_REUSEPORT_BPF)",
> > > +	      "err:%d errno:%d\n", err, errno);
> > > +
> > > +	for (i = 0; i < NR_RESULTS; i++) {
> > > +		err = bpf_map_lookup_elem(result_map, &i, &tmp);
> > > +		CHECK(err == -1, "lookup_elem(result_map)",
> > > +		      "i:%u err:%d errno:%d\n", i, err, errno);
> > > +		nr_run_before += tmp;
> > > +	}
> > > +
> > > +	cli_fd = send_data(type, family, &cmd, sizeof(cmd), PASS);
> > > +	nev = epoll_wait(epfd, &ev, 1, 5);
> > > +	CHECK(nev <= 0, "nev <= 0",
> > > +	      "nev:%d expected:1 type:%d family:%d data:(0, 0)\n",
> > > +	      nev,  type, family);
> > > +
> > > +	for (i = 0; i < NR_RESULTS; i++) {
> > > +		err = bpf_map_lookup_elem(result_map, &i, &tmp);
> > > +		CHECK(err == -1, "lookup_elem(result_map)",
> > > +		      "i:%u err:%d errno:%d\n", i, err, errno);
> > > +		nr_run_after += tmp;
> > > +	}
> > > +
> > > +	CHECK(nr_run_before != nr_run_after,
> > > +	      "nr_run_before != nr_run_after",
> > > +	      "nr_run_before:%u nr_run_after:%u\n",
> > > +	      nr_run_before, nr_run_after);
> > > +
> > > +	printf("OK\n");
> > > +	close(cli_fd);
> > > +}
> > > +
> > >  static void prepare_sk_fds(int type, sa_family_t family, bool inany)
> > >  {
> > >  	const int first = REUSEPORT_ARRAY_SIZE - 1;
> > > @@ -664,6 +712,8 @@ static void test_all(void)
> > >  			test_pass(type, family);
> > >  			test_syncookie(type, family);
> > >  			test_pass_on_err(type, family);
> > > +			/* Must be the last test */
> > > +			test_detach_bpf(type, family);
> > >  
> > >  			cleanup_per_test();
> > >  			printf("\n");
> > > -- 
> > > 2.17.1
> > >